Ask HN: When do we call it engineering and when do we call it research? | Hacker News Reader

Ask HN: When do we call it engineering and when do we call it research?

1 pointVishal191119992 comments

I am confused why ReactJs which introduced important concepts such as VirtualDOM and transformed how we write code for the web, not considered a research project or research work?

Apologies, if my question is too silly,

View on HN

There is not a clear line but I’d say the context matters.

An academic computer scientist produces conference papers about a program they wrote, not a program. Such a person might be wary of doing a live demo at that conference because… things could go wrong. Some CS grad students and profs are brilliant programmers, but not all of them. To write a paper about a program you only need to get the program sorta-kinda working, it doesn’t need to be reliable or have any polish.

A software developer writes programs for a living, it is secondary if they present about them. Such a program has to be good enough for people to really use. The software dev isn’t afraid to do a live demo because they have used the program every day for six months.

In the case of React somebody could have made a prototype and written about it and that would have been research but if you got it packaged up so anybody could use it… that’s engineering.

Even the very quotidian applications programming I do sometimes has an aspect of research; for instance I was working on a query system that is a hybrid of full text and boolean search and discovered a way to write a parser that corrects any mistakes a user types, e.g. if it is

  (A AND B

it gets rewritten to

  A AND B

there was some reading compiler books (without the answer I wanted but a mention of a COBOL compiler from the 1970s that did this) and a lot of thinking about parser theory and drawing illustrations on paper. Out of all that came a product that customers are using everyday.

I was thinking about it this way. That when we have a need to compare the benchmarks to know how good something is, it's a research.

React was launched as a way to write code in a better manner, not in a faster manner, if the former was the case and they compared the performance with something, it might have been a research work, as there were "findings".

What do you think on this?

Reply on news.ycombinator.com